Automated oxygen control for preterm infants receiving continuous positive airway pressure in southwest Nigeria: an open-label, randomised, crossover trial.

Background: Titration of oxygen therapy to target safe oxygen saturation (SpO) values is a vital part of care for preterm infants, but is difficult to achieve, particularly in settings in which oxygen, monitoring technology, and human resources are scarce. We aimed to evaluate the safety and efficacy of automated titration of oxygen therapy partnered with a low-cost continuous positive airway pressure (CPAP) device versus manual oxygen control in preterm infants requiring CPAP in a high-mortality, low-resource setting.

Methods: In this open-label, randomised, crossover trial, preterm infants with a gestational age younger than 34 weeks (or a birthweight <2 kg if gestation was unknown) who were aged 12 h or older and required CPAP and oxygen were recruited at two hospitals in southwest Nigeria. Participants were randomly assigned (1:1) to one of two intervention sequences (ie, to commence the study on automated oxygen control or manual oxygen control) with block randomisation (blocks of 4 and 6) and stratification by health facility. The study statistician was masked to treatment group assignment, but the participants' parents or caregivers and clinical staff were not. Participants received automated or manual oxygen control for two 24-h periods in random sequence. Automated oxygen titration was done with a control algorithm with proven efficacy in high-resource settings. During periods of manual control, oxygen therapy was adjusted by clinicians. The primary outcome was the adjusted mean difference in the proportion of time participants spent in the SpO target range (ie, SpO 91-95% when receiving oxygen or SpO 91-100% when not receiving oxygen) between automated and manual oxygen control, analysed by intention to treat with weighted repeated measures mixed model linear regression. This trial is registered with ClinicalTrials.gov, NCT05508308, and is completed.

Findings: Between Sept 13, 2022, and Sept 11, 2023, 72 infants were screened, and 49 (22 female, 27 male; median gestation 29 weeks [IQR 28-31]; median birthweight 1·2 kg [1·1-1·5]) were enrolled in the study and randomly assigned. A total of 80 study periods for 46 infants contributed data to the analysis of the primary outcome as three (6%) of the 49 participants had no oxygenation data from either study period. The mean proportion of time spent in the SpO target range was higher during automated control periods than during periods of manual control (adjusted mean 88·1% [95% CI 84·0-92·2] vs 30·1% [20·9-39·3]; adjusted mean difference 58·0% [95% CI 48·0-67·9]; p<0·0001). There were no device-related adverse patient outcomes and short-term safety outcomes favoured automated control.

Interpretation: Automated titration of oxygen partnered with a low-cost CPAP device improved time spent in the safe SpO range compared with manual control. There is high potential for this technology to mitigate the risk of oxygen-mediated complications in preterm infants in low-resource settings.

Funding: National Health and Medical Research Council Australia and University of Tasmania.